bloomz
/
evaluation_bloomz
/evaluation_l1
/anli
/dev_r1
/guaranteed_possible_impossible
/results.json
{ | |
"dataset_name": "anli", | |
"dataset_config_name": "dev_r1", | |
"template_name": "guaranteed/possible/impossible", | |
"evaluation": { | |
"accuracy": 0.39 | |
}, | |
"arguments": "Namespace(config_name=None, dataset_config_name='dev_r1', dataset_name='anli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name=None, template_name='guaranteed/possible/impossible', tokenizer_name=None, use_slow_tokenizer=False)" | |
} |